Produced by the Computer Board And Research Councils' Joint Network Team and Network Executive.
Professor (Mike) Wells completed his term as Director of Networking at the end of March. Mike's five-year term has seen many changes including the introduction of the JANET network and the development of an OSI Transition Strategy for the UK Academic Community. There has been a rapid growth in the use of the network infrastructure by all sectors of the Community and the UK's pioneering initiative in the use of open networking has attracted favourable comments from both within the UK and from other countries. I am sure all readers of Network News will join me in thanking Mike for his efforts as Director and wishing him well for the future. Mike will still be involved in the networking programme as a member of the Network Advisory Committee.
The Network Advisory Committee has recommended that the posts of Director of Networking and Head of the JNT and NE should be merged under the title of Director of Networking. As a result of this decision Dr. R (Bob) Cooper has been appointed as the new Director as from April. I am sure Bob requires no introduction to Network News readers and that he will receive your support and best wishes in his new role.
Ted Herbert, Secretary to the Computer Board
In the previous issue of Network News I described the new initiative on international networking and the changes that had taken place with the management and funding of the UK EARN service. During the past year there has also been considerable activity aimed at improving access to the US, particularly access between the JANET and Internet communities.
I am pleased to announce that the US National Science Foundation (NSF) has agreed to fund the US side of a link between JANET and the NSFnet. The NSF approval is in response to a proposal produced by the JNT, UCL, ULCC and the USC ISI (University of Southern California Information Science Institute) in December 1987. The UK side of the project is funded by the Computer Board and the SERC Science, Engineering and Astronomy Boards. Other UK funding bodies have expressed an interest in participating.
NSFnet is planned to be the high performance backbone of the Internet linking the US supercomputer centres and the regional networks and it thus plays a similar role to JANET in the UK. The transatlantic connection will be a 64kbits/sec link between ULCC and the John von Neumann Supercomputer Centre at Princeton using the new TATS transatlantic optical fibre. The initial services offered will be electronic mail, file transfer and terminal access using a clone of the current UCL JANET/ARPA gateway and it is hoped that the new service will remove many of the restrictions and problems that have affected the UCL service over the past few years. It is expected that the new link will enter service by 4Q88 but timescales are dependent on the installation schedule on the US side.
The new service will be managed on the UK side by ULCC under contract to the Network Executive. UCL will be contracted to provide software support. The introduction of the new service will be in stages and as an interim measure it is planned to run the current DARPA access service via the new gateway at ULCC. Progress is well advanced and the ULCC gateway is expected to enter service in June. Part of the development plan will include the provision of improved management facilities including a new mail authorisation mechanism based on trusted hosts. The new mechanisms have still to be finalised and the existing mechanisms will be retained for the new service in the interim.
Agreement has been reached with the MoD for continued access to the Internet via the UCL ARPA link for a six month period prior to the availability of the new transatlantic connection to NSFnet. The use of this link by the academic community is restricted to mail traffic. The cooperation of the MoD in allowing this extended period of use is gratefully acknowledged.
As a result of these developments there will be a significant improvement during 1988 in the network infrastructure linking the UK and US academic and research communities, and reasonable continuity of the existing services prior to the availability of the new facilities.
Bob Cooper
A joint working party (WP-A), set up to define the requirements for an X.25 subnet to serve the European research community, has recently produced an interim report. The working party comprised representatives from RARE, EUNET, EARN, the HEP community and CEC's IES (Information Exchange System) with observers from CEPT. The COSINE Policy Group (CPG) has approved the interim report as the basis for immediate action to allow an initial X.25 service to be available by the start of 1989.
The Dutch PTT has made an offer, via RARE, to provide a pilot service on a suitable timescale within the framework of the emergent Managed Data Network Service Scheme and this initiative has been welcomed by the CPG. The European Commission and the Government of Norway have indicated that they are willing to provide financial support for this project.
These developments represent a significant step in the achievement of an important and long-awaited facility for the European research community. The CPG has stated that it would like to see this infrastructure used by all networks supporting research activities in Europe. Transition of application services should be as smooth and rapid as possible but continuity of services for the end user is an essential requirement.
WP-A has been asked to continue its work towards defining the full requirement and is expected to report again in October.
Bob Cooper
All trunk traffic has been running between the four central JANET switches (Daresbury, London, Manchester and Rutherford) at 256Kbps since the middle of March 1988. We propose to increase the speed of most of these trunk lines to 512Kbps later this year. The next release of the multiplexer software, due around the end of the year, will give us the extra re-configuration capability to increase the speed of the remaining trunk lines.
The first of the site connection upgrades have been completed. 48Kbps connections have been provided to NERC-Swindon, Reading University, Keele University and Heriot-Watt University.
The theoretical maximum packet size available for an X.25 (1980) network is 1024 bytes. The JANET switches are configured to allow the use of 256 byte packets and can pass 512 byte packets. The majority of site upgrades and the high speed trunk lines are connected to DMA driven devices within the switches which are capable of handling large packet sizes. Therefore on selected lines it is possible to increase the packet size available for file transfer type transfers. However the situation with the Campus switches is not at all clear. Some are capable of handling the same size packets as for JANET but in other cases there are restrictions to a maximum of 256 bytes. On some of the switches, any attempt to use packet sizes greater than the value that switch has been configured for will result in that particular link becoming un-available.
Any attempts made to use large (i.e. greater than 256 byte) packets should be referred to the Network Executive so that we can confirm that the proposed route will accept these larger packets.
John Carey
X.400 is the CCITT defined protocol for electronic mail, originally published in 1984. It uses the term Management Domain (MD) for a collection of mail systems operated by an organisation. X.400 defines the interconnection between Administration Management Domains (ADMDs), operated by public telecommunication providers (strictly by RPOAs), and Private Management Domains (PRMDs), operated by anyone else.
After the publication of X.400 by CCITT, ISO started work on extending the definition of X.400 for use between the mail systems, in X.400 terminology Message Transfer Agents, (MTAs), within PRMDs. Unfortunately, ISO's additions made their protocol, called MOTIS, incompatible with strict implementations of X.400(84). CCITT also wanted to extend the functionality of and fix problems in X.400 and collaborative work between ISO and CCITT to produce a joint standard started before ISO had published its X.400(84)-based MOTIS standards. The joint ISO/CCITT standards should be finally ratified by the respective organisations in October and November this year, as ISO 10021 (a seven part International Standard) and by CCITT as X.400(88).
The Academic Community's Transition Report (the White Book) identified the ISO Standard as its target for electronic mail. Apart from the greater maturity of the standard and the fact that it makes proper use of the Presentation Layer, the principal features of most use to the UK Academic Community in ISO 10021 and not contained in X.400(84) are:
Development of a converter between Grey Book Mail and ISO 10021 is already in progress at UCL and due for completion before mid-1989. It is intended that a detailed Operational Requirement for ISO 10021 for the community will be developed this autumn, so that it can be issued (allowing ISO 10021 as an alternative to Grey Book) as soon as MOTIS achieves full International Standard status, provided the mail convertor and NRS OSI developments keep to target timescales.
For communities without the benefit of an existing comprehensive electronic mail network, interim use of X.400(84) makes sense. To enable mail exchange between Grey Book and X.400(84) users, a set of tracking gateways is being provided.
There are files of information about these gateways: use FTP to fetch the file <DOCS>ean.usage from uk.ac.ean-relay for an introduction to the EAN-RELAY service, and fetch <DOCS>x400.usage from uk.ac.ucl.cs for the interim X.400 gateway. In both cases use username guest and your own name as password. There are also electronic mail distribution lists, ean-users and x400-users, giving information on changes, known bugs, etc.
Addressing message recipients is very different to the Grey Book format in ISO 10021/X.400. A structured address called an O/R Address consists of a number of attributes and their values. In order that X.400 users can communicate with Grey Book users, the Grey Book users will need to know how to express their own address in X.400 attributes, and should be encouraged now to start giving their mail address in both formats on, for example, their business cards.
Grey Book systems are strongly encouraged to support the format of local-part (left of the @) consisting of Initials separated by dots followed by Surname which is recommended in Mailgroup Note 5, as this will give their users more logical O/R Addresses.
To obtain the O/R Address for a Grey Book user in UK AC, the first three attributes are:
Then use the Standard Form NRS name to generate:
Then use the local-part (left of @) to form the Personal Name which consists of:
Surname = NTIN36, or Surname = Steve, or Surname = JonesPW.
Some examples may help to illustrate the possibilities and give a suggested layout for business cards etc:
Electronic Mail
Grey Book: J.A.I.Craigie@iik.ac.rutherford
ISO 10021/X.400:
Country GB
ADMD Gold 400
PRMD UK.AC
Organisation Rutherford
Surname Craigie
Initials JAI
Electronic Mail
Grey Book: Black@uk.ac.oxford.physics.vax
ISO 10021 /X.400:
Country GB
ADMD Gold 400
PRMD UK.AC
Organisation Oxford
Org-Unit Physics
Org-Unit Vax
Surname Black
Electronic Mail
Grey Book: Shirley.Wood@uk.ac.nitherford.gec-b
ISO 10021/X.400:
Country GB
ADMD Gold 400
PRMD UK.AC
Organisation Rutherford
Org-Unit GEC-B
Surname Wood
Given-name Shirley
The interim EAN based Grey Book to X.400 gateway at UCL is currently connected to several other PRMDs and to the BT Gold ADMD experimental service. However, most of these PRMDs are serving few users, often just X.400 developers. Two exceptions which are of more general interest are the French Academic X.400 network and the UK Government's internal X.400 system, IDEM (Inter-Departmental Electronic Mail). The current list of PRMDs is available by FTP, fetch the file <DOCS>X400.sites from uk.ac.ud.cs using username guest and your own name as password.
IDEM does not use personal names to address its users, apparently civil servants should be considered merely as Organisational-Units and not as people! However, Grey Book does require something before the @, so the most convenient way to address an IDEM user is to give his personal name even though this is not required or used by IDEM. For example:
J.Craigie@gb.gold~400.idem.ccta.gemini.janet01
The BT Gold ADMD consists, at present of one system running Dialcom software called Mailbridge which gateways from X.400 to other Dialcom systems which provide the Telecom Gold mail service. This system provides relaying of X.400 messages between PRMDs as well as gatewaying to Telecom Gold. At present there are 29 PRMDs connected, of which UK.AC was the first.
Using this ADMD service to relay between PRMDs appears to be satisfactory, although so far little use has been made of this because other PRMDs, to which we don't have direct connections, have few 'real' users. Currently only PRMDs connected directly to the Telecom Gold ADMD are accessible through it; Gold is not yet connected to other ADMDs, which would give connectivity to PRMDs connected to them as well as to their direct users. Some experimental ADMD interconnection is known to be taking place, so connectivity is expected to increase in due course. In fact most other ADMDs are more experimental at present that BT; only the French ADMD, Atlas 400, is offering an equivalent experimental service.
However, attempting to use the gateway to Telecom Gold mailboxes has shown it unsuitable for general user service. It requires both the Gold and X.400 users to be registered in a special directory which translates between the six character Dialcom Mailbox IDs and X.400 O/R addresses, and vice versa. Thus O/R Addresses have to be generated for existing Telecom Gold users and "fictitious" Dialcom IDs created for all external X.400 users, including those gatewayed from Grey Book. There is currently no mechanism for Gold users to get themselves registered in this directory, and the overhead of registering all individual users in our community is unacceptable.
New user-agent software for Dialcom systems, called Mail 400, is currently being tested by Telecom Gold. This provides full support for X.400 O/R Addresses, so when this software is released Telecom Gold users will be able to send and receive messages from X.400 (and hence Grey Book) users without any need for registration. We plan to defer offering a service from the UK.AC to Telecom Gold users until the new Mail 400 software is released to all Telecom Gold users. Unfortunately, BT are unwilling to quote a date for this.
Jim Craigie
In the OSI Transition Report the transition problems of Slotted Rings (better known to some as Cambridge Rings) were not fully addressed, mainly because the ISO standards for this technology were not sufficiently advanced at the time. It was decided to wait until the ISO standards were stable and the response of the manufacturers to the standards became clear. By the end of 1987, ISO had agreed on the text of ISO 8802/7 for Slotted Rings and the JNT convened an ad-hoc group of implementors and users of Slotted Rings conforming to the interim CR82 standard to discuss the future strategy.
The meeting agreed that the best approach was to follow as closely as possible to the recommendations contained in the Pink Book, which describes the use of the ISO Connection Oriented Network Service (CONS) over CSMA/CD LANs. Accordingly, an addendum to the Pink Book was drafted which became known as the "Raspberry Addendum"
However, in subsequent discussions with product suppliers and users it became clear that the existing CR82 hardware, although compatible in terms of the hardware signalling mechanisms with 8802/7. was not suited to the more severe timing constraints of the standard. Following a proposal from CAMTEC Electronics Ltd. the CR82 Users Group asked the JNT to convene a meeting of the CR82 Implementors Group to resolve the problem.
The meeting was held on 24th May 1988 and an acceptable solution was achieved which ties together the elements of the Pink Book, i.e. the use of X.25(84) to provide CONS over the link level control LLC2, and a relaxed specification of 8802/7 which permits the use of existing CR82 hardware (P and R services of the Orange book). This has become known as the "Peach" standard for Slotted Rings.
The change control of this standard has been vested in the CR82 Implementors group, which has resolved not to change it further unless severe and unforeseen technical problems become apparent at the implementation stage.
The JNT now intends to approach existing manufacturers for their plans for products based on the Peach specification. Any such products offered will be able to co-exist with existing CR82 equipment on the same physical medium and it is foreseen that suitable interworking units will be made available. The JNT also intends to subject such equipment to approval tests to ensure adherence to the standard, reliability, and interworking.
The aim of this initiative is to protect the existing investment in Slotted Ring technology while at the same time allowing users and product suppliers to progress towards the use of ISO Standards that are compatible with the rest of the Coloured Book transition.
Willie Black
There is an historical approach to networking which uses the terms LANs and WANs to differentiate between technologies and protocols. Wide Area Networks, WANs, are often related to the slow X.25 packet switching technology due to the early involvement of European PTTs in providing services of this type. This infrastructure was based on low bandwidth links usually associated with the telephone networks. Nevertheless, there emerged a substantial number of local services based on the same cabling and switching technologies.
In the mother country of television, the US, data communication based on cable TV was pioneered using coax cables. Development of other technologies such as Ethernet led to the first LANs with the concept of high speed communication with limited extensibility. Nevertheless, the LANs could be interconnected via dedicated links to form a widely distributed infrastructure. A nation-wide Ethernet based on satellite links was established in the US using this technique.
Like many other organisations the Deutsche Forschungsnetz (DFN) is concerned with the interworking of these two communication cultures, i.e. available X.25 based infrastructure on one side and available LAN-based host manufacturer communications products and technologies on the other.
A valid and possible OSI solution is to interlink LANs via dedicated lines or even X.25 virtual circuits. Nevertheless, such a solution is questionable if:
In the DFN environment these assumptions have been answered by rejecting the historical view of internetting LANs and WANs. In the DFN model the terms LAN and WAN should lose their association with certain protocol stacks or even technologies and stand for local and wide area management aspects only. An overall global network architecture which is as closely related as possible to the widely used public standard seemed to be the more appropriate solution.
As soon as a protocol stack implementing X.25 PLP over LLC Type 2 over CSMA/CD reached the state of an international standard such a solution became very attractive for the OSI-based academic network in West Germany, and since there exists no central procurement policy in Germany to encourage the adoption of standards by computer vendors the project management of DFN was quite happy that the UK JNT was already well advanced in promoting the Pink Book standard and encouraging its implementation for different host systems. The success of the policy in the UK allows us to persuade the computer companies in our country to provide the X.25-based protocol architecture as official products. Thus, in DFN it can be planned to build up a homogeneous X.25 infrastructure with different sub-networks (local X.25 switches, Ethernets, etc.) and to run ISO TPO on top between any DFN-end systems.
But, DFN does not only want to follow in the footprints of the UK academic community but wishes to open some new areas using the Pink Book standard. It is planned, with projects already established, to interlink Ethernet networks via MAC-level bridges and multiplexers using the 140 Mbps overlay network of the Deutsche Bundespost and to interlink this high speed infrastructure on coax and Fibre optics with medium and slow speed infrastructure on copper wires.
The general aim is to achieve a homogeneous network system made up from a diverse technologies but using X.25 as the standard for all access. This will enable us to use the TPO-based functional standards throughout for end-to-end communications.
Thanks to progress in the development of standards we do not need to refer to the UK interim Pink Book standard but can use the European Functional Standard, ENV 41103, which is now available.
Wulf Bauerfeld
This year the fourth international RARE Networkshop took place in Les Diablerets, Switzerland from 16 to 18 May. RARE (Reseaux Associes pour la Recherche Europeenne; the European organisation of national research networks and their users) each year organises a Networkshop in order to have a broad discussion related to the harmonised use of research networks, to review the progress in this area and to stimulate new work within the organisation. To ensure a broad involvement, both geographically and by interest, participation to the Networkshop is by invitation via the RARE members.
Last year's Networkshop was held in Valencia, Spain, and was attended by about 130 people. This year 177 people attended over 40 presentations concerning the various aspects of networking. Some major topics were: the COSINE project, overview of RARE activities, LAN/WAN interworking, future high speed and broadband services, service management, and the various migration paths of existing networks towards OSI.
The Networkshop was hosted by SWITCH, the Swiss research network and CERN, the European institute for High Energy Particle Physics. Prof. Hochstrasser (Swiss Ministry of Education) and Dr. Thresher (CERN) delivered the opening speeches. Prof. Hochstrasser stated that Switzerland wants to contribute as much as possible to the European integration efforts on networking. Dr. Thresher underlined that RARE had achieved much in the last few years. The opening speeches were followed by an overview of the RARE activities, where the RARE working groups delivered their yearly reports.
The presentations were not restricted to the European situation. Zhao Xiaofan (North China Institute of Computing Technology) presented the current networking situation in China, where the use of OSI is required by law (c.f. the European situation for public purchasing following Council Decision 87/95 EEC) and a part of the Chinese 5-year plan. Mr Zhao said: 'China is much more open than ever before and will surely be more open in future, we therefore welcome co-operation' China is considering becoming an Associate Member of RARE. Steve Wolff (National Science Foundation, USA) gave an overview of existing research networks in the US which have come together to form the FRICC (Federal Research Internet Co-ordination Committee), namely NSF, NASA and the federal departments DOE, DOD and DHHS. FRICC hopes to foster better co-ordination between the US and Europe through yet another committee which met just before the Networkshop. Early goals include an OSI-gateway between Europe and the US.
At the Networkshop a plenary discussion was held on the Implementation Phase of the COSINE project. The Eureka project COSINE intends to rapidly establish an OSI environment to make data communication services available to users from academia and industrial research organisations. RARE is currently undertaking the COSINE Specification Phase, which includes the identification of user needs, selection of urgently needed services and if possible future services, evaluating the existing European telecommunication services as well as identifying the range of systems to be provided with communication facilities. The first reports of the COSINE Specification Phase are now ready and will be published shortly. The Specification Phase ended on 1 July 1988. RARE is now in the process of preparing a proposal for the COSINE Implementation Phase, which is hoped to start on 1 January 1989.
RARE's first Annual Report was presented at the Networkshop. This report gives an overview of the main activities of RARE in 1987: the Message Handling Services (MHS) project, the COSINE project, Activities of the RARE Working Groups, the 1987 Networkshop and the Opening of the RARE secretariat in Amsterdam, The Netherlands. An overview of all presentations at the Networkshop will be given in the Proceedings that will be published in a special RARE issue of the Elsevier journal 'Computer Networks and ISDN Systems' in September.
Further information on RARE and copies of the Annual Report can be obtained from the RARE secretariat.
Ruud Olthoff
Due to the anomaly whereby mail to International Grey Book Mail sites is penalised by the difficulty of sending mail through the JANET PSS Gateways, it has been decided that communication between International Grey Book sites and the UK should be encouraged and improved. Therefore International Grey Book sites should register themselves within the NRS with an application relay pointing to the EAN-RELAY. This means that any mail sent from a UK site to an International Grey Book site will be sent via the EAN-RELAY. It should be emphasised that this route will only work for grey book mail sites registered in the NRS. We hope that this will improve communications between the UK and other countries using the Grey Book Standard.
Currently mail via the EAN-RELAY is not charged, but if the traffic through the gateway increases significantly, then some form of charging mechanism may be introduced.
Shirley Wood
If you have a copy of the NRS Lookup Protocol (NLP) specification, you may wish to request a supplement to it containing corrections to minor errors in it, a specification of an additional type of lookup to verify consistency of calling NSAP and SNPA combinations, and an additional protocol allowing re-use of the connection which is intended principally for network service relays and protocol converters. Requests should be sent to M.Thompson@UK.AC.RUTHERFORD.
Three test NLP responder services are now available for NLP initiator implementors to test against. Details are available from Jim Craigie or John Dyer. A general NLP service from the central NRS system will be made available after an adequate amount of controlled testing against other implementations has occurred.
Jim Craigie
The JNT asked the Edinburgh University Computing Service to evaluate the provision of campus NRS servers for the Academic Community. A report has been presented to the JNT and discussed at the NRS Steering Group. The main proposals resulting from the discussion are as follows:-
The complexity of the software dictates that a multi-process operating system be used. It was recognized that either Unix or VMS would be suitable engines but that Unix was preferable on being easily scalable across a range of machine sizes and potential manufacturers.
The name server will be primarily supported on Campus X.25 networks with a full compliment of Coloured Book support, including Red Book. In addition the ability to provide Pink Book access would be a highly desirable feature.
The name server will be implemented on a commercial database. The choice of a database will be made during the implementation phase. The alternatives of either using the existing code of the central NRS service or starting from scratch in a language more suited for Unix, for example C, were rejected on the basis that these would make the eventual package more difficult to maintain and would necessitate the writing of a number of utilities to maintain the database which would be provided as part of the commercial database package.
It is anticipated that an implementation phase will commence in the next two months with the aim of a releasable product in the first quarter of 1989.
Brian Gilmore and George Howat, Edinburgh University Computing Service
JTMP services have been available for some time now at the National Centres, RAL, ULCC and UMRCC. Difficulties with the services have gradually been ironed out, and all three Centres now provide an acceptable level of support.
JTMP provides a standard protocol for job submission, but inevitably leaves some matters of detail open to the choice of the host installation, particularly in the area of naming. Work is in hand in identifying and where possible removing such differences between the National Centre services. Of course the principle view of JTMP for users of the Centres remains their familiar local JTMP user interface, whichever Centre they are accessing.
JTMP is the only job submission protocol which will be supported at all three Centres. Users of the Centres are encouraged to move to using JTMP for remote job access if they are not already doing so.
John Powers on behalf of TRIKE (TRIKE is a meeting of the National Centres Networking groups)
A new electronic mailing list (jtmp-use@UK.AC.RUTHERFORD) has been set up to facilitate discussion of the quality of the job transfer services (especially the server services at RAL, SWURCC, ULCC and UMRCC) currently on offer. This is to be contrasted with, and is intended to complement, the existing mailing list for RBJTMP implementors. It is intended to bring into the discussion User Support staff, as well as Networking people, both at the RBJTMP users' home institution and at the major serving sites.
To join the list, send email to jtmp-use-request@UK.AC.RUTHERFORD.
Phil Jones
At the recent Networkshop 16 in Reading, the format of previous years was modified to make the event more of a workshop than a conference. Additional time was provided for discussion (there were two parallel sessions rather than one) and many of the presentations in plenary were targetted at the discussion sessions which were to follow (rather than having introductions to the topics just before the discussion sessions began).
As in previous years, summaries of the discussion session conclusions have been produced. We felt that these ought to reach a wider audience as quickly as possible, and so we publish here the summaries of five of the six sessions - the summary for the sixth (on Future Performance Requirements and Capabilities) was not available at the time of going to press.
These reports have been grouped as a special pull-out supplement to Network News.
Ian Smith
Chairman - Ian Smith (Network Executive) Reporter - Phil Harrison (University of Nottingham)
This discussion group was attended by about 50 delegates. The discussions covered only the recovery of charges from sites and end-users resulting from use of services/networks that charge on a traffic volume basis. The main focus was on the use of host based applications authenticated on a "trusted NSAP" basis.
At the start of the discussion session the concept of a greengrocers shop model for charging was introduced. This involves separation of the charges levied on the site by the JANET Network Executive from the charges levied on the user by the site management for provision of a service (e.g. sending a mail message, transferring a file, etc). The former charge represents the wholesale costs of buying raw network access in terms of bytes and packets. The latter charge represents the retail cost to the user of using network applications and would be on a fixed scale normally. The site management could choose whether to pass on the cost of "breakages" in the form of failed call attempts, repeated file transfers, or whether to allow for the occurrence of these in the charge for services. It could also choose whether to make a profit in order to cover manpower costs, for example, or whether to subsidise the service for the benefit of all users. The user interface to the network applications at the site would need to provide enough logging information to enable an accounting package to work out the cost associated with the user's activity. Ideally, the interface should tell the user how much the file transfer, mail transfer, etc would cost before committing to it.
Concern was expressed over the amount of local effort required to provide suitable accounting facilities in the existing user interfaces. Not all file transfer implementations provide suitable logging to allow accounting to take place. In general it was not clear how much work might be involved and whether this should be done on a machine range basis rather than on a site by site basis. There was some discussion on the desirability of having a central accounting machine, e.g. for mail.
It was hoped that a portable accounting package would be produced to process the itemised billing that would be made available for each trusted host. It was pointed out that the itemised billing would be on a per-call basis and would include all relevant information for accounting purposes.
It was pointed out that the problems are not specific to the academic community but are equally applicable to large corporations that wish to account for use of chargeable networks by different divisions. There is a lack of management facilities in the ISO standards currently and this situation could continue for some considerable time. The solutions found for the academic community may be applicable elsewhere.
There was considerable discussion on the possibility of user workstations and others masquerading as a different trusted host in order to avoid charges. It is apparently quite easy to masquerade as another host on an ethernet simply by modifying the local software to produce the relevant calling NSAP address. It was thought that the gateway on to the WAN should therefore check the calling MAC address as well as the NSAP for authentication purposes. It was pointed out however that it is possible to change the calling MAC address as well. In general it was agreed that this is probably a local site problem more then a community wide one in that a local department might masquerade as another local department but it would be virtually impossible to masquerade as another site. It was thought that security in general relied heavily on trusting staff and users, that there would always be some security holes, and that other means are available for controlling the extent of abuse e.g. banning users, etc. The greengrocer model for charging applies here in that the retail price of goods includes an element to cover pilferage.
There was some concern about using public terminals and workstations for accessing chargeable services, but it was suggested that these should be treated like telephones, where those in public areas are not given the privilege of making outside calls.
It was hoped that the requirements would be tackled soon in order to gain experience necessary for specifying future products such as interfaces for FTAM and X.400. The distributed directory service could involve lookup charges which are as significant as the charges for the subsequent mail. There may be two charges involved in using a public mail transfer agent, one for the network call to the agent and one for the mail sent by the agent. It is not clear how these charges might be levied.
There was some discussion of existing schemes for local billing to reclaim charges from users. Some used a subscription of around twenty pounds per annum, which made the user think about whether the registration for use of the chargeable service was necessary each year, and also covered the costs of those registrations that were little used. A general
difficulty was perceived over the small bills raised by some users. One site did not issue invoices for amounts less than five pounds. Some sites insisted on a departmental account code being supplied against which the bill could be charged before allowing registration of the user. The subject of VAT was raised. It was thought that this is only a problem for recharging other institutions, internal reclamation of charges was considered exempt providing no extra charges were levied. It was pointed out that people are charged VAT rather than departments. Sites should be encouraged to get the bursar involved in the charging process.
A request was made for a networking ombudsperson to arbitrate in the case of looping mail calls, chronically repeating file transfers, and so on, which could run up large bills. The problem was thought to be soluble on an ad-hoc basis. It was not known how BT would handle such a situation.
The question of terminal access was raised and discussed briefly. It was pointed out that there would not in general be registration of individual users at the central gateway from JANET to the chargeable service. A site would need to provide a local authentication and accounting point through which the call must be made. The address of the accounting point would be registered at the central gateway as the trusted address for all the users registered locally. Alternatively, a local trusted host should be used to make the terminal call. Individual registration would be provided on a minimal basis for sites that could not provide local authentication and accounting facilities. It was agreed that it was not possible to get PAD software modified to provide authentication and accounting facilities.
In conclusion it was felt that there are still a number of difficulties to be resolved and that it may not be possible to satisfy everyone's requirements. However, something must be done soon for the emerging new products. There is probably a significant amount of effort required at each local site to support the model and the costs need to be examined and compared with those for more centralisation.
Chairman - Jim Craigie (Joint Network Team) Reporter - David Roberts ( University of Leicester)
The aim of this session was to discuss the provision within the UK Academic community of a User Directory Service: a computer-based service providing information to humans on how to communicate with other humans. For example, it would enable a user to find another's telephone number or electronic mail address, etc.
It became clear early in the discussion that a genuine need exists for such a service: many of those present reported their experiences of being approached by users who wished to make contact with colleagues working at other institutions. Apart from resorting to the telephone, paper mail or POSTMASTER there is currently no alternative provision in this area.
Although a number of those present wished to implement short-term interim solutions using currently available tools, it was generally accepted that for such a service to be approved and supported centrally it would have to be based on appropriate ISO standards.
Two major themes emerged from the discussion: the mechanism for implementing such a service; and the data content and its administration.
The suggested strategy for implementation was that there should be a distributed solution, based on X.500/ISO 9594, initially mounting one DSA per site. The community should use the experience gained to provide feedback to later revisions of X.500.
The wide spectrum of possible content for such a directory was discussed. This ranges from a simple lookup of usernames for e-mail purposes to a fully-fledged academic information service giving conventional mail addresses, telephone numbers, research interests, papers published, etc.
It was acknowledged that the non-technical problems associated with such data would tend to increase as the breadth of content increased. These problems would include considerations such as privacy, security, sensitivity, and civil liberties aspects.
Possible methods of acquiring and maintaining the data were also discussed. It was concluded that the information had to come from existing sources - the possibility of relying on users to register information about themselves was discounted. Administrative Computing units might in some cases be best placed to maintain site DSAs.
The suggested strategy for data content was that there should be an initial service simply to map personal names onto Grey Book mail addresses or X.400 O/R names. Service providers should be able to make available the data for this purpose. The experience gained should be used to improve methods of data acquisition and to discuss and resolve the 'political' issues outlined above.
A number of sources of influence on the content were put forward. These included:
Several suggestions were made with a view to preventing abuse of the service. These included allowing subjects to be ex-directory, or to blank-out certain fields in their own entry; precluding the ability to generate lists of staff matching certain attributes (network specialists, for example!); and providing a sliding scale of access controls to restrict the availability of certain items of information.
About sixty delegates attended this discussion session, and the impression gained was that a User Directory
is a service which the community needs. Initially, such a service should be mounted simply, both in terms of mechanism and content. The experience gained should then be used to develop towards full academic information services.
A Directory Group has been formed to progress the issues raised in this session, leading to the planned and co-ordinated introduction of Directory Services within the UK Academic Community. Anyone wishing to participate in this group should contact Jim Craigie (J.Craigie@ UK.AC.RUTHERFORD).
Chairman - Phil Jones (Joint Network Team) Reporter - Dave Sedgwick (University of Newcastle upon Tyne)
The aim of the group was to discuss how such tools could significantly improve the provision of information to the user community. The group focussed on how networked and network tools could enable a significant improvement in the information toolset available to people to help them in their work, e.g. in their research and teaching.
Discussion
There was a need to provide significantly improved information to the whole community about the use of JANET and the facilities available over JANET and also information on accessing International facilities. The group agreed that such information services were required on JANET now. However it was agreed that such information was not the prime topic for discussion. This was the tools for information provision. The Network was only one of a range of possible subjects.
The remainder of the discussion was centred around the main points raised in the paper produced by Jill Foster (University of Newcastle).
A national information service should be available as a community facility. A top level server should contain pointers to other information services; it should not be the only information service available. A distributed, hierarchical system of information servers was seen as the best approach. Information about local facilities should be kept on a local information server because it was easier to maintain one copy of a file locally. Obviously the success of this scheme relied on all participants cooperating on the provision of information.
Information Servers should have a user friendly interface, ideally the same user interface, which should be freely available across the community. The service should allow both interrogation by mail and interactive use. A desirable addition to the facility would be FTP. Users of the service should be able to use bulletin boards, list servers, conferencing and digests.
User interfaces and query languages were not discussed in detail by the group, but obviously these were areas which required investigation and possibly standardisation.
A number of similar systems were already in existence around the world and the group stressed the need to investigate these systems before deciding on a final design for an Academic community service.
The group realised that funding was required to set up and operate information services and looked to the UGC or the EEC for some of the funding. If a good information server was produced it should be marketed to encourage the rest of the community to adopt the same tools for building other information services. Currently NISS are in the process of installing a bulletin board information service, using the HUMBUL software, on a MicroVax. This project has funding for 3 years.
The group concluded the following:
Chairman - Willie Black (Joint Network Team) Reporter - Peter Barry ( University of Glasgow)
The group discussed the following as options for use in file access:
DAP: Although this is a DEC proprietary product, the specifications are available. However as it is both proprietary and not available on a sufficiently wide range of machines, it was discarded as a candidate for an open standard.
NFS: This de facto standard runs over UDP/IP connectionless lower layer architecture and is therefore incompatible in its current form with the UK Academic Community's connection-oriented architecture based on international standards. It is however available now on a wide range of machines.
FTAM: Although the standard, and profiles for file transfer are available today, neither the products nor profiles for file access exist as yet. Nevertheless in the long term it appears to meet all of the requirements for file access and will be implemented as a connection-oriented application.
There is a wide spectrum of styles of file access. It was agreed that bootstrapping, paging, swapping, etc should not be done over either JANET or even over a Campus network. The requirements for disk access rather than file or data access could only be supported by lightweight non-OSI protocols over a localised LAN, e.g. NFS.
The following points were made
The following areas should be studied, and the JNT may fund work in these areas.
Chairman /Reporter - Mike Wells (University of Leeds)
The discussion concentrated on three main topics:
The classes of establishment eligible for connection to JANET are enshrined in the list of eleven types of institution whose connection can be approved. Unfortunately this list contains a number of phrases which admit of a variety of interpretations, of which the most tricky is the last, catch-all, class "Such other institutions as .... may decide from time to time"!
In general terms, commercial establishments should only use any connection to JANET for communication with non-commercial establishments. Their initial connection may in some cases be directly to a JANET switching centre, or may be indirect via a connection to a JANET site. There is a long paper, which originally started life as a short note, which attempts to clarify the position, and which will appear in a future issue of University Computing.
There was discussion of the extent to which individual sites are required to install controls (i.e. software) to limit use by non-academic establishments. In general terms such controls are not necessary, although if they can be applied they may be. It is sufficient to advise any non-academic sites of the limitations as to use (see (1) above). If misuse is detected, the correct course is to tell the offending person to stop, and to advise him that failure to stop may result in much more serious action. The law accepts that it may be difficult to impose controls; however, a persistent misuse is an offence, and may be treated as such by the courts. To date, very few cases of warnings being issued have arisen; in every such case the warning has been heeded, usually with an apology. In theory, the law does not recognise ignorance as a valid defence. In practice ignorance in this particular area is accepted as a valid excuse for previous misuse.
Many sites are presently planning to replace their speech systems. The question of shared use of installations was discussed. The law requires all parts of a speech system which connects to a PTO (i.e. to British Telecom or to Mercury) to be maintained by an appropriately qualified person. The maintainer may insist that all installation intended for speech are kept totally separate from that used for data; this is a decision by the maintainer, not a legal requirement. In practice, it appears that it is often feasible to share cable ducts, cable trays etc, but that using some pairs in a multi-pair cable for speech and other pairs for data will not be acceptable.
In February of this year, the final GOSIP V3.0 documentation was published. Demand for the documentation has been consistently high and has met with a warm response from all major suppliers, departments and other specialist organisations who have received copies.
The Central Computing and Telecommunications Agency (CCTA) has had a long- standing commitment to OSI, it has, however, at the same time acknowledged that unless standards specialists are available, a purchaser may experience great difficulty in specifying standards when purchasing communications based products. To help overcome these difficulties, the GOSIP (Government Open Systems Interconnection Profile) Project was launched some two years ago. The project's major thrust was the creation of a specification that would meet the needs of an administrative environment as well as being capable of interfacing directly with the procurement process.
The most significant feature of the GOSIP documentation is the use of a standard form known as a PICS - Protocol Implementation Conformance Statement - which contains a definition of requirements, derived from the early drafts of the international PICS, for a particular sub-profile within GOSIP The user is not obliged to make any additions to this proforma: the usual proliferation of choices within standards has been reduced and guidance provided on how to specify the remaining options where these exist. This can then be made contractually binding for the purchaser, supplier and tester.
Testing authorities were recognised to be key players in the process, and were included, along with representatives from departments and suppliers, in the project steering committee and task forces. CCTA is currently investigating the possibility of some kind of fixed price 'insurance' based tests being made available to purchasers at the acceptance stage of a contract, although in the initial stages of procurement GOSIP recommends suppliers' own tests as being sufficient.
The GOSIP documentation is of special interest to public bodies; since February of this year, public authorities in Europe are bound by the EEC Decision 87/95/EEC which makes it mandatory for purchases over 100,000 ECUs (approximately 71,000 pounds) to be OSI based. This Decision is expected to have a major impact on the marketplace for OSI products as the public sector in the UK and Europe begin to demand a wider range of products to meet their obligations. CCTA have produced a report which provides information on the current products and plans for OSI of 24 of the major IT suppliers. This report, together with the GOSIP documentation which is already in use in Departments will be of significant value to project officers embarking on OSI procurements.
Mark Gladwyn
The following information is up to date at the time of going to press. If you are working on SSMP software not mentioned below, the Joint Network Team would be interested to hear from you. Contact John Dyer.
SSMP terminal emulators provide the means to communicate between an asynchronous terminal or PC to an SSMP application running in a host. In the case of an asynchronous terminal, this is achieved by the use of some add-on hardware, eg the Fawn Box. In the case of a PC, software is installed either as Firmware or using a floppy disc. Arrangements have been made for copies of the disc based versions of SSMP emulators for a range of PC's to be distributed to all Academic Computer Centre Directors. You should therefore be able to obtain a copy of many of the emulators listed from your Computer Centre. In case of difficulty contact John Dyer at the JNT
1. The Fawn Box provides SSMP facilities in conjunction with an existing VDU terminal. In addition to SSMP, it provides either a null emulation, VT100-like emulation, or a conversational buffer. The following terminal types are supported:
DEC VT100, VT320; 41 xx series;
Wyse 75; Tektronix
Televideo 9220, 912 (and DT2, DT22), 910 (and 925);
Hazeltine 1500 (and 1510, 1420), Esprit (models I and II);
Cifer Systems 2605, 2634;
BT Merlin; Lear-Siegler ADM3A; Dacoll 242/1
The Fawn Box is based on small microcomputer systems from Oval Automation and Sherwood Data Systems. Both versions come boxed (to stand on a desktop) with mains lead and plug. The JNT arranged a one-off special purchase at a price of £127 each inc. VAT. Some of these units are still available. Further information from, and orders to John Dyer at the JNT.
2. BBC Micro SSMP version 2.00, for BBC Models B, B-f and Master. Obtainable from Graham Dixon, Churchill College, Cambridge CB3 ODS. Cost £25 + VAT (one off), site discounts available. Graham's e-mail address is WGD1@UK.AC.CAMBRJDGE. PHOENIX.
3. IBM PC (MSDOS) and clones. The NUMAC SOFT Terminal Program, version 2.22. Written by Clive Gerrard and Dave Sedgwick of the University of Newcastle upon Tyne. Provides a conversational buffer or VT1OO-like interface, with SSMP available from either. Obtainable free of charge (via FTP) from the University of Lancaster microcomputer software archive. Contact
Dave_Sedgwick@UK.AC.NEWCASTLE for details.
4. IBM PC (MSDOS) and clones. KERMIT with SSMP (KERMITS). SSMP support written by Brian Holley of the University of Cambridge. Obtainable free (via FTP). Contact the author: BJH6@UK.AC.CAMBRIDGE.PHOENIX.
5. Research Machines NIMBUS PC. The NUMAC SOFT Terminal program (see (3) above). Obtainable cost £10 pounds inc. VAT, from Brian Jones, Computing Laboratory, University of Newcastle upon Tyne, Newcastle NE1 7RU.
6. ACT Apricot PC. The NUMAC SOFT Terminal program. This version is no longer being developed, but the current (that is, the last) version is shipped with the Nimbus PC version described above.
7. Atari 1040ST and 520ST. Based on the NUMAC SOFT Terminal program. Operates in high resolution monochrome and medium resolution colour. Contact Brian Robertson ( B.Robenson@UK.AC.ABERDEEN)
8. DEC Rainbow. Again a port of the Newcastle SOFT v2.1 terminal program undertaken by Aberdeen University. Maintenance and distribution is now being handled by Newcastle who should be contacted with any enquiries.
9. Acorn 32016 second processors running the Panos operating system via a version of Cambridge's Terminal and Transfer Program (TTP). Contact Philip Hazel (PH10@ UK.AC.CAMBRIDGE.PHOENIX)
10. Acorn Archimedes, also via TTP, available when serial port problem resolved. Contact Philip Hazel, as above.
1. 3270 Terminal Emulator version 1.1.2. Author Pat Moran of the University of Edinburgh. Provides 3270 emulation via SSMP from VM/CMS systems. Available from IBM, contact your local sales representative or Neil Dunlop of IBM Cambridge.
2. CURLEW, a portable full-screen editor. Author Alan Hunter of the University of Newcastle upon Tyne. Available for EMAS, VAX/VMS, UNIX 4.3bsd, VM/CMS (from line mode), Honeywell CP-6, MTS. Public domain software.
For the VAX/VMS version contact Alistair Scolley (AJSCV@UK.AC.ST1RLING.VAXA).
For UNIX systems contact Chris Ritson (Chris_Ritson@UK.AC.NEWCASTLE).
For availability of GEC OS4000 and Primes versions contact John Dyer at the JNT.
For general queries contact Alan_Hunter@UK.AC.NEWCASTLE.
Local (non-network) versions for Sun-3, IBM PC and RM Nimbus PC are under test.
3. SHED, the Sheffield Screen Editor. Author Peter Mason of the University of Sheffield. Provides an SSMP full-screen editor for PRIMOS. Available from Peter Mason (CSJPM@ UK. AC.SHEFFIELD.PRIME A).
4. E, a portable full-screen editor. Author Philip Hazel of the University of Cambridge. Available for Phoenix/MVS and VAX/VMS. The editor has also been implemented on the Acorn 32016 and Archimedes. Contact Philip Hazel (PH1()@UK.AC.CAMBR1DGE.PHOENIX) for more details.
5. SECCE, the Edinburgh Compatible Context Editor converted for SSMP with a Curlew-like full-screen editing mode. Full-screen conversion by Keith Yarwood of the University of Edinburgh. Initially available for EMAS. Contact K.Yarwood@UK.AC.EDINBURGH for further details.
Consideration is also being given to a proposal to mount the ECCE screen interface on MVS/Phoenix at ULCC
6. SEDIT, the EMAS line mode editor converted as for SECCE. Other information as for SECCE.
7. SVIEW, a full-screen document display system. Another Keith Yarwood conversion, this time of the Edinburgh VIEW system which allows full-screen display of either flat files or, with a suitable markup, structured documents. In addition to standalone usage, on EMAS SVIEW provides online help for CURLEW, SMAIL, etc.
8. SMAIL, a full screen interface to the EMAS Mail system. Contact Keith Yarwood as above.
9. SCALENDAR, a full-screen calendar display utility, by Keith Yarwood of Edinburgh University, for EMAS. By default shows last, current and next month's calendar entries, but can be steered to other months either by use of the cursor keys or by entering a month and year.
Page 14
10. SCOMPARE, a new full-screen text file comparison program for EMAS. Contact Keith Yarwood as above.
For anyone thinking of writing an SSMP host application, or converting an existing one for SSMP, Newcastle University can make available a set of Pascal routines which provide an SSMP host application virtual terminal interface. Included with them is a program, AVOCET, which demonstrates many of the facilities of SSMP. Contact Alan_Hunter@UK.AC.NEWCASTLE for details.
In 1985 the Computer Centre and the Department of Sociology and Anthropology at University College of Swansea were jointly awarded a grant from the "Computers in Teaching Initiative" to study the operational and sociological implications of placing computing equipment in student residential areas.
Very early in the project, it became clear that key factors in running a successful remote service would be the ability to monitor the state of communications equipment, and to place news messages in X25 Camtec PADs. A monitor was therefore designed and commissioned with these aims, and has since become a vital tool in network management for both remote and local equipment. The monitor is based on a humble BBC microcomputer, and displays the number of calls through each PAD, its up-time, and (since we are currently upgrading to release 4 software) also flags PADs which are still running release 2.5 software. Calls are also made to hosts, primarily to check that all communication links are operational. The code will fit into EPROM, so allowing use with a disc-less BBC.
The success of the monitor has led to a more ambitious project to archive the configuration of an X25 Camtec PAD to BBC disc, and restore it should the PAD subsequently fail. The archive operates by scanning the replies to commands such as "SHOW LINE 3" for interesting strings, and storing sufficient information to enable the program on the restore phase to answer all questions relating to lines, address lists, groups and text. Additionally, the PAD prompt and PAD identifier are set corresponding to the names known to the monitor.
The archive/restore suite has an extra utility which is the archive editor. This very useful tool will enable us to synthesize an archive for one PAD from that of another, so that new PADs can be configured
automatically provided that we have another in some location which offers the same user interface. All necessary parameter changes, such as calling addresses, are translated.
The tool will also transform an archive file taken from a PAD running release 2.5 software into one suitable for restoring to a release 4 PAD. This facility is proving extremely useful in our upgrade programme for all PADs, and to date seventeen have been upgraded. Our upgrade mechanism involves archiving the PAD in question over the network during normal user service, running the edit program, plugging the Beeb into port 0 of a spare PAD with release 4 EPROMS in order to reprogram it directly, then swapping boards with the final target system. Despite the widely different management interfaces between the two releases, the task is often performed by operators with absolutely no expertise in communications and proves a very effective and painless way of managing the upgrade with little PAD downtime and no problems with even the more complex PAD configurations.
In view of the initial funding for this project, we are prepared to make the above tools available at no cost to bona-fide JANET sites via electronic mail. Support will, necessarily, be somewhat less than open-ended. but the tools are in service in Swansea and in general it would be in our own interests to rectify any shortcomings. Please send mail requests to CCBRADY@UK.AC.SWANSEA.VAX
Phil Brady and Tony Oilier, Computer Centre, University College of Swansea.
Richard Startup, Dept of Sociology and Anthropology, University College of Swansea.
Electronic distribution lists provide an efficient means of sharing knowledge, asking and answering queries within a User community.
Many such lists exist, but typically only a small percentage of the relevant community knows of them. This note discusses reasons for this and of an initiative to help solve the problem.
Electronic distribution lists (DLs) have been around for some time. The JANET network is one reason that they are becoming more common. Electronic mail is an every day fact of life for many researchers, so it is natural to see e-mail as an important means of spreading and sharing knowledge.
DLs provide an economic way forward. There is still one major problem: How does one become aware of DLs in existence? Typically a list is announced publicly once and then disappears. How do new members of a community find out about these lists ?
The JANET news machine is one, limited, possibility. But even knowledge of the news machine is restricted.
After a small amount of research the problem seemed to be both widespread, and genuine. I started to ask whether anybody knew of a list of lists, nobody did -but most people contacted said that they would like one and encouraged me to create one.
To date I have discovered about 216 lists, including a contact point for each list. I publish a monthly list detailing all these lists. Perhaps the most interesting list, with respect to this article, is lists@UK.CO.IMPER1AL-SOFTWARE-TECH. This list is used to mail out copies of the list of lists.
To keep this list up to date I need to receive details of as many lists as possible. Specifically: the name and purpose of the list, the electronic addresses for submissions and requests. By convention requests to the list administrator are usually sent to list-request.
Neil Todd
In order to improve information dissemination and discussion about CSMA/CD open systems networking (i.e. OSI and Coloured Books) an electronic mail distribution list is being established for the academic community. In the first instance one or two people only per site will be put on the list, usually from the Computer Centre, together with technical experts that are members of the JNT LAN Advisory Group. This is necessary in order to keep the list to a manageable size. It is expected to have the distribution list in operation some time in July and an existing paper mailing list maintained by the JNT is being used as the basis for establishing membership of the electronic mail version, The new membership list will be given in an e-mail message; if any site believes that there are errors or omissions they should send a message to jnt-b@UK.AC.RUTHERFORD. In particular if you do not receive any message to your site by the end of July, please contact jnt-b@UK.AC,RUTHERFORD.
In order to cater for the wider audience, there will be an article giving an update on CSMA/CD activities in the next issue of Network News since a number of products, key components, host interfaces etc, are currently under test and an extensive programme of host interoperability testing is also in progress.
Les Clyne
At large computing service installations, there will normally be a formal method of dealing with problems. An incident which represents the occurrence of, for example, a serious but not catastrophic fault, perhaps due to a bug for which the system supplier is responsible, might go through the following process over a period of time:
There will be (at least) three reference labels created during this sequence of events:
Support for smaller systems or indeed for third party software, may be less formal, but there will be analogies. The supplier should still have the formal aspects of this in place.
That much is, of course, not unique to communications problems - far from it. However, one of the topics that comes up in liaison between the JNT (and the Network Executive) and suppliers is Problem Management and bug clearance. The JNT may, on behalf of the community, press a supplier to deal with a particular bug as a matter of significantly greater urgency, for example. Alternatively, the JNT may be consulted about an order of priority for fixing bugs or dealing with design faults.
Many readers will be at sites where problem management is systematic, but most will agree, one imagines, that there are those who could be more systematic. It would certainly help the JNT to help the community if this were so. For example:
Phil Jones
Fibre Distributed Data Interface (FDDI) standards are being progressed through the relevant ISO committees. These should offer Local Area Networks based on token ring technology, running on fibre optic cables, at around 100 Mbit/s. Such LANs could act as Campus backbones to which other, slower speed, technologies could connect at the MAC layer. Alternatively, hosts could connect directly to the FDDI network and obtain high-speed host-host throughput.
Several Universities have expressed interest in installing FDDI equipment. However the standards are not yet fully agreed at the ISO level and the implications of rushing immediately into this technology are uncertain.
At this year's Networkshop, the JNT proposed the setting up of an FDDI Advisory Group and invited interested technical experts to join. The first aim of the group will be to further define its terms of reference and to set limits on what should/can be achieved. Perhaps it could arrange for one or more pilot projects to begin to evaluate the technology on behalf of the whole Academic Community. Certainly it should come to some conclusion on the functional standards to be adopted as Computer Board policy in this area.
The first meeting of the group has not yet been arranged and I am still collecting the names of interested people.
Willie Black
There is a growing interest in dial-up modems which have error detection and correction capabilities. We are aware that several sites have evaluated and are using these modems. If you have some experience of one or more products that are currently available please could you let us know your conclusions so that we can try and provide helpful information to others. Please contact John Carey of the Network Executive.
Bob Cooper
I have been persuaded that it is a good idea to have a mailing list to cover matters of interest common to both Unix and open networking (i.e. not Unix specific), or closely related matters for which there is no other suitable forum. The address of the list is unix-comms@UK.AC.RUTHERFORD
Requests for additions are welcome. These, and other enquiries about the list should be sent to unix-comms-request@UK.AC.RUTHERFORD
It would be helpful if people would try not to send messages to this list where there is an existing proper alternative. For information on other lists, send a message to lists-request@UK.CO.IMPERIAL-SOFTWARE-TECH
Phil Jones
Networkshop 17 is to be held at Warwick University from 11 to 13 April 1989. The format will be similar to Networkshop 16, however the programme has yet to be decided. If there is a networking issue that you would like to see discussed which would be of general interest to those working in the Academic Community, please submit it to the organisers for consideration.
The organisers are:
Tim Clark at Warwick, John Dyer for the JNT